臺灣博碩士論文加值系統

這篇論文針對嵌入式應用程式的關鍵執行路徑提出一個創新的能源管理方法。在動態電壓調整技術中，作業系統排程利用行程的執行時間資訊來執行動態電壓調整。但是排程程式在程式執行時無法得知程式內部中確實的時間資訊。在我們的論文，我們使用電源提示將程式內部的時間資訊傳達給作業系程排程程式。我們藉由剖析程式來找出即時系統應用程式中的能量關鍵執行路徑，並根據能量值的資訊安插能量提示在這些路徑中以減少能源消耗。接下來，電源提示包含在最差情況中的剩餘執行循環數，此資訊可以被即時作業系統使用在動態排程。當作業系統排程者處率能量提示時會有負荷。因此，我們安插能量提示在能量關鍵路徑上以減少非必要的提示數目。我們的實做是基於MachSUIF編譯器架構，Wattch模擬器，以及作業系統排程模擬程式，並使用MPEG2解碼程式做效能評估。在我們的實驗裡中，可以節省25% 的能源，並且只有0.5% 的負擔。

This paper presents an innovative power management technique focusing
on the critical path for real-time embedded applications. We ¯rst ¯nd the
hot spot in real-time application by pro¯ling the programs and insert the
power hints in it based on the energy cost information to reduce the energy
dissipation. Next, the power hints that convey the remaining execution cycles
in the worst case will be used by the real-time operating system to schedule
the tasks at run time.There are overheads when OS scheduler computing
power hints. So we insert power hints in the energy hot paths to reduce
the number of hints which is unnecessary. Our work is implemented on
the basis of MachSUIF compiler infrastructure, the Wattach simulator, and
the OS scheduling simulating program and measured by using MPEG2 as
benchmarks. On average, the experimental results show that our work can
save the energy by 25% and with overheads 0.5%.

1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Background and Related Work 4
2.1 DVS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1.1 Compiler . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1.2 Real-Time Operating System Scheduler . . . . . . . . . 6
2.2 SUIF and MachSUIF Compiler System . . . . . . . . . . . . . 6
2.2.1 SUIF . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2.2 MachSUIF . . . . . . . . . . . . . . . . . . . . . . . . 7
3 Algorithm 10
3.1 System Architecture . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 Pro¯ling Critical Paths . . . . . . . . . . . . . . . . . . . . . 12
3.3 Inserting Power Hints . . . . . . . . . . . . . . . . . . . . . . 19
4 Experiments 23
4.1 Experiment Environment . . . . . . . . . . . . . . . . . . . . 23
4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5 Conclusion and Future Work 26
5.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Bibliography 26

[1] N. AbouGhazaleh, B. Childers, D. Moss, R. Melhem, and Matthew Craven.
Energy management for real-time embedded applications with compiler sup-
port. ACM SIGPLAN Languages, Compilers,and Tools for Embedded Sys-
tems, 2003.
[2] N. AbouGhazaleh, D. Moss, B. Childers, R. Melhem, and Matthew Craven.
Collaborative operating system and compiler power management for real-
time applications. IEEE Real-Time Embedded Technology and Applications
Symposium, 2003.
[3] A. Azevedo, I. Issenin, and R. Cornea. Pro¯le-based dynamic voltage schedul-
ing using program checkpoints. In: Carlos K, ed. Proc. of the Conf. on Design,
Automation and Test in Europe. Washington: IEEE Computer Society Press,
pages 1{8, 2002.
[4] Stanford Compiler Group. The SUIF Compiler Infrastructure. Stanford Com-
piler Group, Stanford, March 1995.
[5] C.H. Hsu, M. Hsiao, and U. Kremer. Compiler-directed dynamic frequency
and voltage scheduling. In Workshop on Power-Aware Computer Systems,
November 2000.
[6] Chung-Hsing Hsu, Ulrich Kremer, and Michael S. Hsiao. Compiler-directed
dynamic voltage/frequency scheduling for energy reduction in mircoproces-
sors. International Symposium on Low Power Electronics and Design 2001,
pages 275{278, 2001.
[7] Intel, inc. Intel Pentium M Processor. http://www.intel.com/products/ pro-
cessor/pentiumm.
[8] Intel, inc. Intel XScale Technology. http://www.intel.com/design/intelxscale.
[9] C.M. Krishna and Y.-H. Lee. Voltage-clock-scaling adaptive scheduling tech-
niques for low power in hard real-time systems. In Proceedings of the 6th Real
Time Technology and Applications Symposium (RTAS 00), May 2000.
[10] A. Manzak and C. Chakrabarti. Variable voltage task scheduling for minimiz-
ing energy or minimizing power. In Proceeding of the International Conference
on Acoustics, Speech and Signal Processing, June 2000.
[11] T. Pering, T. Burd, and R. Brodersen. Dynamic Voltage Scaling and the
Design of a Low-Power Microprocessor System. In Power Driven Microarchi-
tecture Workshop, attached to ISCA98, June 1998.
[12] P.Pillai and K. G. Shin. Real-Time Dynamic Voltage Scaling for Low-Power
Embedded Operating Systems. In Proceedings of the 18th ACM Symposium
on Operating Systems Principles, pages 89{102, October 2001.
[13] D. Shin and J. Kim. Intra-task voltage scheduling on dvs-enabled hard real-
time systems. IEEE Design and Test of Computers, March 2001.
[14] D. Shin and J. Kim. A pro¯le-based energy-e±cient intra-task voltage schedul-
ing algorithm for real-time applications. International Symposium on Low
Power Electronics and Design, pages 271{274, 2001.
[15] Y. Shin, K. Choi, and T. Sakurai. Power Optimization of Real-Time Embed-
ded Systems on Variable Speed Processors. In Proceedings of the International
Conference on Computer-Aided Design, pages 365{368, November 2000.
[16] Michael D. Smith. Machine SUIF Compiler. Division of of Engineering and
Applied Science, Harvard University, March 1998.
[17] Transmeta, inc. Crusoe Processor Technology. http://www.transmeta.com/
crusoe/index.html.